Learn how AI can enhance security measures by detecting malicious assets, and safeguarding against vulnerabilities. Stay secure with advanced technology.
The internet is huge and is expanding every day. While this has many positives for businesses, managing the potential risks in this environment can be daunting. The reality is that across the internet brands are susceptible to everything from brand abuse to phishing websites. So, how is it possible that organisations can keep track of the web-based assets that belong to them, and at the same time distinguish them from something that may only look like it’s theirs? Finding and verifying all of a company’s web assets across the entire internet is a massive undertaking. You essentially need to filter the whole internet and try to pick out what is relevant, and then set about detecting the risks – or even potential risks – within what you have found.
This isn’t a process that can be managed manually. The staff-hours alone would make this hugely prohibitive, and that’s without taking into account the potential margin for error. Instead, it requires a different approach, one based around automation. At Darktrace, my team and I work on exactly those kind of solutions. We’ve developed our own algorithms to define what distinguishes a client’s brand-owned site from everything else on the internet. We refer to that as a company’s Brand DNA. These special characteristics help us to predict and identify where any brand-related assets are across the entire internet, and how they should be investigated further. The concept of an organisation’s Brand DNA breaks down into two areas: what is unique by design and what is uniqueby comparison.
Unique by Design
All brands have different design elements that help set them apart from other brands. This can be everything from their name and logo, to the different fonts and colours they use in all their communications and websites. By ingesting this data into our algorithms we are able to scan the internet for any web-based assets that may be relevant to a company, based on these key brand elements. While the elements of ‘unique by design’ are relatively easily understandable by humans, no organisations want to have their time taken up manually searching through millions of images every day in order to help them locate the web properties that might belong to their organisation.
Unique by Comparison
Conversely, ‘unique by comparison’ focuses more on the elements that a human would probably not be able to figure out by themselves. As we suggest potential new domains to customers, we build up a pool of assets. Automation allows us to find patterns in these assets that might not be immediately obvious to humans, such as elements of metadata,nameserver details, or even where a website is hosted. Although unique by design is more about what you actually see on a website and unique by comparison focuses on the back-end, in reality there are overlaps and the two things feed into each other.
As a very basic example: if a domain is hosted on nameserver where other company assets are hosted, AND the company logo is on the page, then the chances are this domain is a company-owned asset. In effect, the two approaches strengthen each other. By analysing all these details together, our algorithms can increasingly accurately score how likely an asset is to be owned by a company. I should add here that unique by comparison is based on comparing a lot of features at once, so it is often not as clear cut as the above example.
Combining Humans and AI
Ultimately, automating the process in this way helps to create a minimal-touch process for companies. The algorithms do all the filtering, enabling the creation of a much-reduced list of assets for the company to look through. Basically, we’re able to break down that list to avery small percentage of the internet that they actually need to look at and then analyse the risk those assets poseto the organisation.
We also use something which we internally label “AI2”(artificial intelligence with analyst interaction). This essentially means we’re adding a human layer to the automated process, for both input and output checks. While the algorithms do all the heavy lifting and aid scalability, the human element allows us to finetune or dive deeper into certain automated findings.
Detecting Malicious Assets
While the algorithms are principally focused on establishing a brand’s attack surface, a useful byproduct is that they can also locate malicious assets, such as potential phishing sites. For example, if something looks like it belongs to the customer, but doesn’t actually belong inside their directdigital infrastructure, then clearly there is an increased likelihood that it is either a brand abuse or phishing site.
On top of this, as part of our search process we can also automatically create combinations of possible URLs that cover common search errors such as typos or “fat finger”errors within brand names, and then hunt for those – clearly the likelihood of these URLs being rogue sites is greatly enhanced.
The Clearest View of the Attack Surface
By combining all these elements, companies are able to get the most complete view of their potential attack surface.And with the use of enhanced automation techniques they can do so with minimum effort. From this position companies are able to easily and quickly home in on the genuine items, and the areas that pose them the most risk. They can then use the resulting list to form the foundations from which they can apply the rest of their security strategy.
Like this and want more?
Receive the latest blog in your inbox
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
NEWSLETTER
Like this and want more?
Stay up to date on the latest industry news and insights.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.
What began as a method to achieve unauthorized access to an account, often driven by the curiosity of individual attackers, credentials theft become a key tactic for malicious actors and groups, as stolen login credentials can be abused to gain unauthorized access to accounts and systems. This access can be leveraged to carry out malicious activities such as data exfiltration, fraud, espionage and malware deployment.
It is therefore no surprise that the number of dark web marketplaces selling privileged credentials has increased in recent years, making it easier for malicious actors to monetize stolen credentials [1]. This, in turn, has created new opportunities for threat actors to use increasingly sophisticated tactics such as phishing, social engineering and credential stuffing in their attacks, targeting individuals, organizations and government entities alike [1].
Credential theft example
TA577 Threat Actor
TA577 is a threat actor known to leverage stolen credentials, also known as Hive0118 [2], an initial access broker (IAB) group that was previously known for delivering malicious payloads [2]. On March 4, 2024, Proofpoint reported evidence of TA577 using a new attack chain with a different aim in mind: stealing NT LAN Manager (NTLM) hashes that can be used to authenticate to systems without needing to know plaintext passwords [3].
How does TA577 steal credentials?
Proofpoint reported that this new attack chain, which was first observed on February 26 and 27, was made up of two distinct campaigns. The first campaign consisted of a phishing attack featuring tens of thousands of emails targeting hundreds of organizations globally [3]. These phishing emails often appeared as replies to previous messages (thread hijacking) and contained zipped HTML attachments that each contained a unique file hash, customized for each recipient [3]. These attached files also contained a HTTP Meta refresh function, which triggered an automatic connection to a text file hosted on external IP addresses running as SMB servers [3].
When attempting to access the text file, the server requires an SMB session authentication via NTLM. This session is initiated when a client sends an ‘SMB_COM_NEGOTIATE’ request to the server, which answers with a ‘SMB_COM_NEGOTIATE’ response.
The client then proceeds to send a ‘SMB_COM_SESSION_SETUP_ANDX’ request to start the SMB session setup process, which includes initiating the NTLM authentication process. The server responds with an ‘SMB_COM_SESSION_SETUP_ANDX’ response, which includes an NTLM challenge message [6].
The client can then use the challenge message and its own credentials to generate a response by hashing its password using an NTLM hash algorithm. The response is sent to the server in an ‘SMB_COM_SESSION_SETUP_ANDX’ request. The server validates the response and, if the authentication is successful, the server answers with a final ‘SMB_COM_SESSION_SETUP_ANDX’ response, which completes the session setup process and allows the client to access the file listed on the server [6].
What is the goal of threat actor TA577?
As no malware delivery was detected during these sessions, researchers have suggested that the aim of TA577 was not to deliver malware, but rather to take advantage of the NTLMV2 challenge/response to steal NTLM authentication hashes [3] [4]. Hashes stolen by attackers can be exploited in pass-the-hash attacks to authenticate to a remote server or service [4]. They can also be used for offline password cracking which, if successful, could be utilized to escalate privileges or perform lateral movement through a target network [4]. Under certain circumstances, these hashes could also permit malicious actors to hijack accounts, access sensitive information and evade security products [4].
The open-source toolkit Impacket, which includes modules for password cracking [5] and which can be identified by the default NTLM server challenge “aaaaaaaaaaaaaaaa”[3], was observed during the SMB sessions. This indicates that TA577 actor aim to use stolen credentials for password cracking and pass-the-hash attacks.
TA577 has previously been associated with Black Basta ransomware infections and Qbot, and has been observed delivering various payloads including IcedID, SystemBC, SmokeLoader, Ursnif, and Cobalt Strike [2].This change in tactic to follow the current trend of credential theft may indicate that not only are TA577 actors aware of which methods are most effective in the current threat landscape, but they also have monetary and time resources needed to create new methods to bypass existing detection tools [3].
Darktrace’s Coverage of TA577 Activity
On February 26 and 26, coinciding with the campaign activity reported by Proofpoint, Darktrace/Email™ observed a surge of inbound emails from numerous suspicious domains targeting multiple customer environments. These emails consistently included zip files with seemingly randomly generated names, containing HTLM content and links to an unusual external IP address [3].
Figure 1: A summary of anomaly indicators seen for a campaign email sent by TA577, as detected by Darktrace/Email.
Figure 2: Details of the name and size of the .zip file attached to a campaign email, along with the Darktrace/Email model alerts triggered by the email.
The URL of these links contained an unusually named .txt file, which corresponds with Proofpoint reports of the automatic connection to a text file hosted on an external SMB server made when the attachment is opened [3].
Figure 3: A link to a rare external IP address seen within a campaign email, containing an unusually named .txt file.
Darktrace identified devices on multiple customer networks connecting to external SMB servers via the SMB protocol. It understood this activity was suspicious as the SMB protocol is typically reserved for internal connections and the endpoint in question had never previously been observed on the network.
Figure 4: The Event Log of a ‘Compliance / External Windows Communication’ model alert showing a connection to an external SMB server on destination port 445.
Figure 5: External Sites Summary highlighting the rarity of the external SMB server.
Figure 6: External Sites Summary highlightin that the SMB server is geolocated in Moldova.
During these connections, Darktrace observed multiple devices establishing an SMB session to this server via a NTLM challenge/response, representing the potential theft of the credentials used in this session. During this session, some devices also attempted to access an unusually named .txt file, further indicating that the affected devices were trying to access the .txt file hosted on external SMB servers [3].
Packet captures (PCAPs) of these sessions show the default NTLM server challenge, indicating the use of Impacket, suggesting that the captured NTLM hashes were to be used for password cracking or pass-the-hash-attacks [3]
Figure 7: PCAP analysis showing usage of the default NTLM server challenge associated with Impacket.
Conclusions
Ultimately, Darktrace’s suite of products effectively detected and alerted for multiple aspects of the TA577 attack chain and NTLM hash data theft activity across its customer base. Darktrace/Email was able to uncover the inbound phishing emails that served as the initial access vector for TA577 actors, while Darktrace DETECT identified the subsequent external connections to unusual external locations and suspicious SMB sessions.
Furthermore, Darktrace’s anomaly-based approach enabled it to detect suspicious TA577 activity across the customer base on February 26 and 27, prior to Proofpoint’s report on their new attack chain. This showcases Darktrace’s ability to identify emerging threats based on the subtle deviations in a compromised device’s behavior, rather than relying on a static list of indicators of compromise (IoCs) or ‘known bads’.
This approach allows Darktrace to remain one step ahead of increasingly adaptive threat actors, providing organizations and their security teams with a robust AI-driven solution able to safeguard their networks in an ever-evolving threat landscape.
Credit to Charlotte Thompson, Cyber Analyst, Anna Gilbertson, Cyber Analyst.
Credential Phishing: Common attack methods and defense strategies
08
Jul 2024
Credential theft remains a top cybersecurity threat
Adversaries have many options in their arsenal to gain access into an organization.
Exploitable vulnerabilities: This can provide access into a system’s processes and allow activity within the context of the service account.
Weak or misconfigured systems: These can provide direct avenues of access into exposed systems.
However, the more desirable option is to obtain user or API credentials permitting the adversary to authenticate and operate as one of the organization’s authorized entities.
While 2023 noted a marked increase in vulnerability exploits as the chosen vector of attack, the use of credentials by adversaries still ranked #1 at 24% in the latest Verizon Data Breach Investigations Report. Mandiant’s M-Trends report noted 14% of their investigations involved stolen credentials as the attack vector, and Darktrace’s 2023 End of Year Threat Report revealed that Credential Access was one of the most observed MITRE ATT&CK tactics.
Credential phishing methods
There are many ways an adversary can obtain a user’s credentials. Some require gaining access to the target system or exploiting an application while others target the end-user directly.
Social Engineering: Many users have a habit of incorporating things in their life into their passwords. Family members, important dates, hobbies, movies, and music favorites have all been used. Adversaries know this and will scour social media to gain knowledge about their intended target. This method was beautifully demonstrated in the 1983 movie, Wargames, where Matthew Broderick’s character scours articles, papers, and video about Dr. Stephen Falken, finally guessing that the password into the WOPR (War Operations Plan Response) computer is that of his deceased child, Joshua.
Credential Cracking / Dumping: If the adversary has gained access to a targeted system, they may employ a password cracking, or credential dumping, program. For Unix-based solutions, obtaining the /etc/passwd and /etc/shadow files provides the users, groups, and encrypted passwords. Adversaries can exfiltrate these files and then utilize password crackers such as John the Ripper, Crack, or codebreaker003. Mimikatz(see more below) can also pass cache information for Mac / Unix and Linux systems.
Windows-based solutions: Adversaries have successfully utilized programs such as Mimikatz to dump credentials and hashes. Mimikatz can pass the hash string to the Local Security Authority Subsystem Service (LSASS) to authorize user actions, as well as perform “kerberoasting”. Kerberos is how Windows systems authorize users utilizing a 3-entity authentication method and symmetric key cryptography to create “tickets” that authorize requested actions. Mimikatz can use Kerberos tickets to gain non-expiring domain administration credentials (Golden Tickets) or tickets to login as a service on the network (Silver Tickets).
Post-It Notes: As organizations and applications started requiring stronger passwords that met complexity requirements, users did what you would expect to ensure they didn’t forget them. They wrote them down (this was also demonstrated in Wargames). The modern-day equivalent is to create a text file with all your passwords (or API credentials) in it – something adversaries are delighted to find.
One of the funniest, yet totally on-point, comic routines I’ve seen on this topic is Michael McIntyre’s You Should Probably Change Your Password skit at the London Palladium.
Phishing / Smishing: Forged messages requesting users to reset their passwords or directing them to enter their credentials used to be easier to spot. However, the emergence of Artificial Intelligence (AI) is allowing adversaries to create very realistic messages and web pages that mimic an organization’s authentication pages. These attempts are not just limited to email, adversaries are utilizing SMS messages and other collaborative communication solutions like Microsoft Teams to transmit fake messages to unsuspecting users. Also, security teams are seeing increased use of Quick Response (QR) codes in scam messages. QR codes are appearing in all aspects of everyday life (I’m finding it hard to go into a restaurant without having to scan a QR code to read the menu) and there is a false sense of security people have in thinking that QR codes are safe to scan.
Vulnerability Exploits: Gaining access to the credential cache or password file is not the only way adversaries can obtain user credentials. Some applications will store the user credentials in process memory (decrypted). If the application is vulnerable to a remote exploit, it can be possible for the adversary to dump the memory of the application process and locate these stored credentials. This was clearly illustrated in the Heartbleed exploit disclosed to the public in 2014.
Air Cracking: Air Cracking is specific to Wi-Fi networks and involves cracking programs that analyze wireless encrypted packets and extracting WEP or WPA/WPA2 PSK passwords (giving the adversary access to the Wi-Fi network).
Dark Web Purchase: Threat groups know how to monetize compromised credentials. Selling compromised credentials on the Dark Web occurs on a regular basis. Sites such as HaveIBeenPwned.com can assist users in determining if a particular password has been found to be compromised. Note: Users should ensure that the sites they are checking to see if their password has been compromised are actual legitimate sites and not a credential harvesting site!
What is credential stuffing and why is it so effective?
Credential Stuffing is so successful because users tend to utilize the same, or very similar, passwords across all the systems and applications they access. This includes both personal and business accounts. Once an adversary harvests credentials from one site, they will try that password on other sites, and if that fails, they can utilize generative AI to predict potential variations of the password.
How to reduce the risk of credential stuffing?
Users can help reduce exposure of their credentials by creating passwords that meet complexity requirements but are also easy to remember. A good approach is to take a phrase and apply a substitution rule. For example, let’s take the start of Charles Dicken’s book A Tale of Two Cities and create a substitution rule for it:
It was the best of times, it was the worst of times
Let’s shorten that to: Best of times Worst of times
Apply the following substitution rule: o = 0, i = 1, e = 3, spaces = @
Now my phrase becomes: B3st@0f@t1m3s@W0rst@0f@t1m3s
You now have a 28-character password that contains letters, a capital letter, number, and special character. Nobody is cracking that, and the phrase and substitution rule makes it much easier to remember (PS: 12-character passwords are also fine, taking ~34,000 years to crack using current technology).
Organizations can reduce exposure through implementation of two-factor authentication (2FA), so even if the passwords are compromised through the methods described above, another authentication layer stands in the way of the adversary.
Additionally, preventing phishing messages from landing in user’s inboxes (Email or collaborative solutions such as Microsoft Teams) is critical not only for reducing the potential exposure of user credentials, but also user’s opening malicious attachments or links. Generative AI tools such as ChatGPT have resulted in over an 135% increase in novel social engineering attacks.
How Darktrace protects against sophisticated credential phishing attempts
Malicious actors can exploit these leaked credentials to drastically lower the barrier to entry associated with brute-forcing access to their target networks. While implementing well-configured MFA and enforcing regular password changes can help protect organizations, these measures alone may not be enough to fully negate the advantage attackers gain with stolen credentials.
In early 2024, one Darktrace customer was compromised by a malicious actor after their internal credentials had been leaked on the dark web. Subsequent attack phases were detected by Darktrace/Network and the customer was alerted to the suspicious activity via the Proactive Threat Notification (PTN) service, following an investigation by Darktrace’s Security Operation Center (SOC).
Darktrace detected a device on the network of a customer in the US carrying out a string of anomalous activity indicative of network compromise. The device was observed using a new service account to authenticate to a Virtual Private Network (VPN) server, before proceeding to perform a range of suspicious activity including internal reconnaissance and lateral movement.
Unfortunately for the customer in this case, Darktrace’s autonomous response was not enabled on the network at the time of the attack. Had it been active, it would have been able to autonomously act against the malicious activity by disabling users, strategically blocking suspicious connections and limiting devices to their expected patterns of activity.
Adversaries have various methods available to compromise user and API credentials. There is no single silver bullet that will protect users and organizations, but rather, a layered approach that incorporates education, security controls such as 2FA, unsupervised AI to detect novel and sophisticated spear-phishing messages, as well as protection against exploits that give adversaries access to systems.
